Journal of the American Medical Informatics Association Volume 12 Number 1 Jan / Feb 2005 13

Viewpoint Paper j

Accelerating U.S. EHR Adoption: How to Get There From Here.

BLACKFORD MIDDLETON, MD, MPH, MSC, W. ED HAMMOND, PHD,

A b s t r a c t Despite growing support for the adoption of electronic health records (EHR) to improve U.S.
healthcare delivery, EHR adoption in the United States is slow to date due to a fundamental failure of the healthcare
information technology marketplace. Reasons for the slow adoption of healthcare information technology include
a misalignment of incentives, limited purchasing power among providers, variability in the viability of EHR products
and companies, and limited demonstrated value of EHRs in practice. At the 2004 American College of Medical
Informatics (ACMI) Retreat, attendees discussed the current state of EHR adoption in this country and identified steps
that could be taken to stimulate adoption. In this paper, based upon the ACMI retreat, and building upon the
experiences of the authors developing EHR in academic and commercial settings we identify a set of recommendations
to stimulate adoption of EHR, including financial incentives, promotion of EHR standards, enabling policy, and
educational, marketing, and supporting activities for both the provider community and healthcare consumers.
j J Am Med Inform Assoc. 2005;12:13–19. DOI 10.1197/jamia.M1669.

The annual meeting of the American College of Medical basic product features, and the rapid cycle turnover of HIT
Informatics (ACMI) in February 2004 focused on the status companies to date. We identify four broad areas for action
of electronic health records in the United States. Attendees to stimulate U.S. EHR adoption. They are: (1) financial incen-
at the meeting discussed three questions regarding the state tives to stimulate the EHR marketplace, (2) EHR functional
of electronic health record (EHR) adoption in this country: and related informatics standards setting and adoption, (3)
Where have we come from? Where are we today and why? enabling policy for EHR adoption, and (4) educational, mar-
And with widespread adoption of EHR in the United States keting, and supporting activities. Before we discuss each of
in mind—How do we get there from here? These discussions these areas in turn, we start by first assessing the current mar-
have been synthesized into three companion reports address- ket and business case for HIT in the United States.
ing each question in turn.1,2 In this report, building on the first
two, we review discussions and recommendations that focus Addressing a U.S. Health Care Information
on the third question. This report represents the opinions of Technology Market Failure
the authors, informed by the ACMI discussions, and does We believe four principal reasons explain the U.S. HIT market
not represent a consensus statement from ACMI. failure. These issues must be overcome to facilitate rapid EHR
We believe that U.S. health care information technology adoption in this country.
adoption is stymied by a fundamental health care informa-
tion technology* (HIT) market failure. The HIT market has HIT Value Proposition
failed because of misaligned incentives (asymmetric risk While a great deal of work has been done demonstrating the
and reward) among key market players, the inability to impact of clinical information systems on clinical decision
achieve broad standards adoption and lack of definition of making and the quality of care, little work has been done
that demonstrates the impact of health care information tech-
nology on economic outcomes. Several studies3,4,5,6,7 suggest
that there should be a positive long-term return on invest-
Affiliations of the authors: Clinical Informatics Research & De-
ment for EHR in the ambulatory care environment and a busi-
velopment, Center for IT Leadership, Partners HealthCare System
and Harvard Medical School, Boston, MA (BM); Department of ness case for standardized interoperability between EHR
Community and Family Medicine and Department of Biomedical implementations,8 but there is limited solid evidence demon-
Engineering, Duke University, Durham, NC (WEH); School of strating significantly improved financial outcomes resulting
Nursing and College of Engineering, University of Wisconsin- from HIT investments. In the absence of solid economic evi-
Madison, Madison, WI (PFB); Center for Biomedical Informatics, dence for EHR adoption, most technology vendors make
University of Pittsburgh, Pittsburgh, PA (GFC).
Correspondence and reprints: Blackford Middleton, MD, MPH, MSc,
Clinical Informatics Research & Development, Partners HealthCare,
*We define ‘‘health care information technology’’ broadly as
93 Worcester Street, PO Box 81905, Wellesley, MA 02481; e-mail:
including in patient and out patient care settings clinical information
<bmiddleton1@partners.org>.
management systems used by clinicians and ancillary staff for the
Received for publication: 08/09/04; accepted for publication: purpose of clinical information management, order entry, documen-
09/22/04. tation of care services, and decision support.
14 MIDDLETON ET AL., Accelerating U.S. EHR Adoption

a business case focused on reduced costs associated with in- Recent research suggests there is a considerable societal bene-
formation management, potential quality of care impact, fit—a U.S. savings potential of $78 billion annually—that
and, to a lesser degree, care process efficiency. For health could be achieved with seamless, fully interoperable health
care providers who heretofore have borne little risk for the care information exchange among key stakeholders in the
quality of care provided, and who from their perspective health care delivery system.2 At the local level, however, cur-
are managing clinical practice well with a paper-based med- rently there is no financial reward for improved clinical infor-
ical record, such arguments have largely fallen on deaf ears, mation exchange among health care entities that regularly
perhaps appropriately so. act as business partners providing care to a common set of
We believe that the research agenda should focus on the value patients—providers are not reimbursed for electronic informa-
proposition of EHR across the health care delivery spectrum. tion exchange. In one case of federally mandated interop-
Solid evidence of economic impact of EHR will help make the erability standards, the Health Insurance Portability and
business case for EHR and drive adoption. Accountability Act (HIPAA) engendered cooperation among
a set of diverse and potentially competing entities to improve
Misaligned Incentives reimbursement transactions and administrative information
In addition to the absence of solid evidence on the economic exchange12 through collaboration to identify and implement
impact of EHR, certain analyses suggest that the HIT market a shared methodology for administrative transaction manage-
is failing due to a fundamental misalignment of incentives be- ment. This idiosyncratic approach, however, is unlikely to lead
tween providers purchasing HIT and those who fund health to a coordinated set of standards adopted across the country.
care, such as public and private payers and employers.1 The In the current marketplace, in the absence of a similar shared
economic analysis suggests that the benefits of HIT do not ac- and realizable gain for clinical information exchange, or other
crue to those who must invest in these technologies. For ex- recognition of the value of collaboration, there is no incentive
ample, many of the patient safety and quality effects of from the individual provider’s perspective for the adoption
EHRs accrue benefit to the payer or employer–purchaser of and use of a common set of interoperability standards. Viewed
health care services who is at greater risk for a patient’s total from another perspective, by distributing the costs of poor
health care costs given decreasing rates of provider reim- information exchange and interoperability far and wide
bursement under capitation. Under fee-for-service reimburse- across all participants in the health care delivery system,
ment models, providers have little incentive to use EHRs each individual entity may be acting rationally from a local
unless they can contribute enough to practice efficiency or perspective, but no entity perceives the magnitude of the
revenue cycle management to improve net revenue per time lost value in the aggregate (A. Milstein, personal communi-
unit. Under mixed reimbursement models such as variable cation, July 2004). This behavior precludes spending by indi-
withholds, and newer pay-for-performance programs, EHRs vidual providers or purchasers of HIT for a potential public
may contribute to achieving performance or quality bench- good dependent upon the cooperation of other independent
marks that warrant increased reimbursement or increased entities. When the vendors of HIT do not perceive their cus-
return of withhold payments. We identify recommendations tomers stating interoperability as a requirement of their sys-
that may stimulate EHR adoption through reimbursement tems, they act rationally and do not include these features
reform and improved capital availability for the provider in their products. Thus, there is an opportunity for a third
sector. party, such as the federal government and private payers,
to introduce reimbursement or regulatory policy changes
Standards Adoption that would drive standards adoption.
Another component of the market failure we identify is de-
layed standards adoption. In the absence of a clear business Viable Companies and Products
case or value proposition, it is not surprising that voluntary The difficulties we have described prevent the development
U.S. standards-setting efforts have made slow progress in of a robust marketplace for HIT. While the academic literature
stimulating adoption of standards addressing HIT functional- has produced solid evidence on the impact of HIT on patient
ity, interoperability, content representation, and messaging. safety,13 the quality of care,14,15 care process efficiencies,16 and
With the exception of recent U.S. federal efforts at standards even revenue cycle management,17 the absence of a clear busi-
acceleration and identification and endorsement of a select set ness case and fundamental misalignment of incentives has
of standards for use in federal programs,9 progress has been protracted the emergence of this market. To its credit, the ven-
slow in the private sector among HIT vendor companies, ture capital community recognizes the potential value of HIT
with the notable exception of the recent effort of the Markle and has made considerable investments over the last decade;
Foundation and the Connecting for Health Program10 and however, the market remains characterized by a few large
the Integrating the Healthcare Enterprise (IHE) efforts.11 vendors (typically with diversified product portfolios, not
The absence of a solid business case for interoperability al- solely dependent upon their HIT product lines) selling to pro-
lows vendors to take a myopic view of the use of standards viders with sufficient operating margins and capital reserves
in their products and in installations of their technology in to make considerable investments and stay the course, and
customer environments. Few HIT customers currently pro- a large number of small, highly unstable smaller EHR ven-
pose information exchange with clinical business partners dors with a relatively short mean lifespan. These same pro-
as a requirement for their clinical systems. Commonly today, viders are typically at risk themselves, through self-insured
only messaging standards are applied to ensure information or publicly funded health plans, for a percentage or all of their
exchange between disparate systems within the context of health care expenditures and thus experience the rewards of
a single health care entity, whether group practice, hospital, HIT investments themselves through internally aligned in-
or multi-facility integrated delivery network (IDN). centives (for example, Kaiser Permanente, Veteran’s Health
Journal of the American Medical Informatics Association Volume 12 Number 1 Jan / Feb 2005 15

Administration). The majority of physicians’ office environ- HIT, this would have the likely effect of causing providers to
ments and small and midsize community hospital settings adopt HIT to capture relevant performance measures and
have yet to make significant HIT investments and in many produce timely reports, warrant the bonus payment, or re-
cases do not believe they are in a financial position to do so. turn withheld reimbursement from payers. The costs of pro-
In the absence of a viable marketplace, and with little barrier ducing such benchmark reports and gathering the requisite
to entry in the absence of a standards conformance require- data without HIT would be greater than doing so with HIT.
ment, the HIT industry is replete with hundreds of EHR Such an approach imposes fewer constraints on the providers
vendors attempting to provide products to fulfill niche re- in terms of what constitutes acceptable HIT, and thus many
quirements from just a few customers, paying little attention secondary benefits may not as readily accrue—either to the
to functional, data representation, or interoperability stan- providers or to society. Nevertheless, this may be the least
dards for EHR. While the EHR may be conceived as a ‘‘system invasive mechanism that could leverage much of the existing
of systems’’18 the lack of clarity around basic product defini- HIT and produce dramatic results. The Bridges to Excellence
tion, relevant standards, and market segments, stifles de- program19 and the Leapfrog Group standards20 are notable
mand from a wary customer. examples of such efforts for out-patient and in-patient care
settings, respectively, and several similar programs are in de-
velopment or pilot stages around the country. Nevertheless,
Recommendations there is a paucity of evidence on what the effect of specific fi-
We now discuss four areas in which we suggest action is war- nancial incentive mechanisms on EHR adoption is—this
ranted to help stimulate the adoption of EHRs in this country: should be a research priority and tops our list of recommen-
market incentives; EHR and informatics standards; enabling dations (Table 1).
policy; and educational, marketing, and supporting activities.
We prioritize these in a manner that we believe will lead to the
quickest response, and in several areas efforts are underway. Table 1 j Recommendations to Stimulate U.S. EHR
However, some require additional effort, and we suggest Adoption
these initiatives may proceed concurrently. This report repre-
Expand the HIT Research Agenda
sents the authors’ opinions on these issues informed by the
1. Increase funding to evaluate the impact of HIT in practice, with
ACMI meeting held in February 2004, evidence in the litera- a focus on economic outcomes, costs and benefits.
ture, and experiences in both academic and commercial set- 2. Evaluate the utility of ‘‘open source’’ or public domain software
tings developing health care information technology. for EHR and implementation and maintenance methods for such
systems.
Market Incentives Financial Incentives to Stimulate EHR Marketplace
Given the heterogeneity of the U.S. delivery system and reim- 1. Reimbursement reform: Establish financial incentives for the use
bursement mechanisms, we recommend the use of market of EHR in practice.
mechanisms to stimulate HIT adoption. Such mechanisms 2. Capital availability: Establish low-interest loans or a grant pro-
take essentially one of two forms and can be expected to pro- gram to facilitate hardware and software adoption in health care
duce increased EHR adoption. We also suggest that an EHR settings.
certification process is called for to identify use of HIT that 3. EHR Certification and Accreditation: Establish a process to certify
EHR products as having requisite functionality in accordance
warrants one or both of these mechanisms. Finally, we sug-
with accepted standards and an accreditation process for level of
gest that open source technologies may have a role in lower-
use of EHR in practice.
ing the price of HIT applications or components (particularly HIT Standard Setting
knowledge components), or pieces of the technology infra- 1. Coordinate existing efforts to specify essential standards for basic
structure for interoperability. We discuss these recommenda- EHR functionality, data representation, and messaging.
tions in turn. 2. Specify a minimal clinical data set covering a patient’s de-
Reimbursement Reform mographics, medications, medical conditions, allergies, advance
directives, and selected data pertinent to patient safety and health
The most direct way to stimulate any market is to increase de-
care quality.
mand. Such an increase would occur if users of HIT were di-
3. Specify minimal functional standards for HIT systems in acute
rectly or indirectly rewarded for using HIT. A direct reward care and inpatient care settings, personal health records, and key
could arise if, for example, payers required not only submis- functional components such as CPOE.
sion of administrative claims data electronically, but also sub- Enabling Policy
mission of any attendant clinical information or other claims 1. Promulgate Medicare Modernization Act relaxations to Social
attachments in electronic form, and a differential payment Security Act, Sec. 1877 (Stark).
was made to the provider supplying these data. This require- 2. Establish federal policy on clinical data ownership and steward-
ment would stimulate the adoption of HIT so that providers ship.
could supply both clinical and administrative data electroni- 3. Establish policy framework for Regional Health Care Information
Authorities.
cally. Secondary uses of these data for both individual pro-
4. Establish U.S. national licensure in the health professions.
vider performance assessment and secondary population
Educational, Marketing, and Supporting Activities.
health surveillance and public health management would 1. Establish educational and marketing campaign for the public—
produce secondary gains. ‘‘Got EHR?’’.
A more indirect mechanism would be to reward providers for 2. Establish educational campaign for health professionals.
attaining desired performance benchmarks across a variety of 3. Establish educational campaign for health care management.
acute and chronic care conditions in both in-patient and out- 4. Create a National Health Care Information Technology Resource
Center.
patient care settings. While not a direct reward for the use of
16 MIDDLETON ET AL., Accelerating U.S. EHR Adoption

Capital Availability the end user that should be open source, but rather the en-
We believe that reimbursement reform is only one part of abling technology and knowledge infrastructure underlying
what will necessarily be a two-part approach to market in- and supporting the end-user application. Much like the U.S.
centives. While reimbursement reform may cause an increase interstate highway system was viewed as a critical infrastruc-
in operating revenue for providers adopting HIT, it does not ture for any form of transportation, public or private, we
address the fundamental capital barrier that providers face in suggest that there are analogous critical information infra-
making the initial investment and capital outlay in expensive structure components undergirding local HIT applications
HIT software, hardware, support services, and the lost reve- used in offices and hospitals that will enable the National
nue typically associated with HIT adoption in the early phase Healthcare Information Infrastructure (NHII). These may in-
of implementation. Many experts have described and pro- clude regional transaction hubs or information exchanges,
posed a variety of mechanisms to increase capital availability secure networks and patient-matching infrastructure, pub-
to small office environments, community hospitals, and lic-interest organizational structures to manage regional
other care settings, that lack sufficient capital reserves or information exchanges and broker communitywide invest-
credit to access capital markets. We suggest that what has oc- ment and serve as a local certification authority, and so on.
curred in many other countries should occur in the United In addition, currently, each provider organization wrestles
States: low interest loans or even one-time grants to pro- with the task of implementing and maintaining knowledge-
viders adopting HIT are in the nation’s and the payers’ inter- based rules and alerts in its HIT applications. This time-con-
est to catalyze HIT adoption. Whatever the approach to suming and difficult task could be ameliorated if there were
providing initial hardware and software it must include an accessible library of such knowledge in the public domain.
a workable plan to both sustain and update those systems. If such component tools and technologies were open source
We need to recognize that commitment to HIT is not a one- and readily available in the public domain, it could have
time expense. a profound impact on vendors building HIT technologies as
it would reduce their internal development costs and mitigate
EHR Certification Process risks of adopting standards.
We believe that an EHR certification process is called for to at-
test to the appropriate functionality of EHRs, and an accred-
itation program is called for to attest to the level of use of an EHR and Informatics Standards
EHR in practice. HIT purchasers need assurance that their A great deal of activity in recent years brought considerable
technology purchases will warrant incremental payments attention to the issue of standards development and accelera-
from payers. Payers need to be assured that their incentives tion of this process.4,8,22 Yet, from either a public or private
are going to physicians who are using more than a spreadsheet perspective, adoption of even a minimal set of standards re-
as their EHR—it must meet minimal functional standards. In mains rare with a few notable exceptions.23,24 The standards
addition, they need to have assurance that the system is being development organizations have focused primarily on spe-
used appropriately to achieve patient safety and quality gains; cific standards such as messaging and have assumed that
for example, that each clinician is using an electronic prescrib- other groups would develop the additional necessary stan-
ing module for every prescription. While this may be viewed dards for complete interoperability including terminology
by some as something that raises the bar for entry into the HIT standards and a reference information model. More recently,
marketplace, this concern pales in comparison to fears HIT HL-7, for example, has begun to develop standards for
purchasers have that their investments will be for naught or broader areas, addressing the complete set of standards that
concerns of the payer community, who fears being asked to is necessary for interoperable data exchange. Lack of aware-
comply with one or more of the reimbursement mechanisms ness regarding existing standards, confusion about which
we have described with no means to ensure compliance standard is the right standard, and lack of proof of the value
with HIT adoption or adequacy of the HIT itself. of standards has severely limited the adoption and imple-
mentation of standards.
Open Source EHR and Related Technologies We believe that specification of a minimal set of essential
Finally, another market mechanism to stimulate market de- standards that have the property of supporting interoperabil-
mand for a desired product is to lower the price. Many pun- ity (the ability to exchange clinical information reliably) is
dits have written about so-called ‘‘open source’’ software critical to rapid adoption of HIT—and a key component in de-
systems in health care, and many providers describe their riving value from HIT.7 It is beyond the scope of this report to
willingness to pay for HIT at a price-point that is far below recommend specific standards, but we suggest that efforts un-
current prices for EHR.21 While it may be debated whether derway in the Consolidated Healthcare Informatics initiative,
the open source model may ever truly apply to EHR applica- the newly created Commission on Systemic Interoperability,
tions given their complexity, rich knowledge content for deci- and private sector efforts at HL-7 and ASTM, be coordinated
sion support, and mission-critical nature, what is clear is that to ensure successful definition of essential standards for clinical
for many providers, one of the main barriers to adoption is information content representation and messaging.25
the cost of current technology. However, it is important to Beyond the specification of standards for clinical information
note what has been successful using the open source process. content and messaging, additional work is needed in specify-
Most successes have been with tools and technology compo- ing a variety of uniform clinical information data sets to facil-
nents rather than large applications. itate interoperability between EHR implementations. The
The question of open-source software and component tech- Continuity of Care Record effort26 is a notable example in
nologies warrants critical analysis and may be addressed at which a set of information is defined to facilitate transfer of
many levels: perhaps it is not the EHR application before patients between health care entities for care and is a useful
Journal of the American Medical Informatics Association Volume 12 Number 1 Jan / Feb 2005 17

intermediate step toward seamless health care information Before discussing regional information exchange, however, it
exchange and interoperability. We believe such instruments is useful to address clinical information ownership and stew-
should be based upon a minimal set of patient care informa- ardship. Many physicians express concern about adopting
tion that includes patient demographics, insurance coverage, HIT when they cannot be assured that the information will
allergies, medications, current medical problems and condi- be made available to them should they elect to switch EHR
tions, and the patient’s advance directives. Such a core data vendors. In addition, de-identified, aggregated clinical data
set serves as a means by which clinicians may quickly become may be viewed as a critical public good in light of bioterror-
familiar with a patient and serves as a foundation for clinical ism and protecting the public health—biosurveillance and epi-
decision support in electronic health records. Availability of demiology research would be well served through access to
a common core set of laboratory data, and such ancillary in- anonymous clinical data arising from EHRs. The regulations
formation as prior electrocardiogram, would also be useful implementing the HIPAA provide guidance for managing in-
to promote patient safety and health care quality, and reduce formation security and privacy. These guidelines have been
redundant utilization. used effectively to facilitate the most notable demonstration
Another area requiring definition and clarity is in the area of of clinical information exchange to date—the Indiana Net-
functionality of clinical information systems. Not only is this work for Patient Care (INPC). In this case, clinical informa-
useful from the business perspectives described above, it also tion is shared broadly across the greater Indianapolis
is critical for enabling the interoperability of an essential min- metropolitan area. Stewardship for the data is provided by
imal care data set and is essential for certification purposes the Regenstrief Institute, which is well versed in clinical infor-
that will warrant additional payments or other incentives to mation management and has the leadership, technical capac-
providers from payers when the use of an EHR can be docu- ity, and political capital to help establish policies and
mented and attested to. The HL-7 functional model of the procedures for the INPC. The lessons learned from this dem-
EHR is an excellent start, and the draft standard is now avail- onstration and others5,28 should be collected and elevated to
able for trial use.27 It is clear, however, that much more work the national policy level so that other communities wishing
needs to be done on functional standards for personal health to create similar regional health care information exchanges
records that interact with EHR systems, inpatient clinical in- could readily adopt policies and procedures that work.
formation systems, and additional detail and specification re- While the HIPAA legislation, including the Privacy Rule, es-
garding critical functional modules such as provider order tablished protections for the security and confidentiality of
entry and clinical decision support. personally identifiable health care information, it does not
Enabling Policy address fundamental issues of data ownership. Clarification
We identify four areas in which national policy could have of the rights of both the providers who gather and collect pa-
a profound impact on the adoption of HIT: modification of tient data, and the patient as source of the data, would be use-
Stark antitrust regulations, policies to guide clinical data ful to help establish the value of these data and appropriate
ownership and stewardship, mechanisms to support creation uses of the data in exchange for compensation in research
of regional health care information authorities, and lastly, and and marketing purposes. Clarification of these rights and
with a longer view of clinical practice in this country, estab- privileges will help define the methods to obtain patient con-
lishing means for national professional licensure in the health sent and grant access to or exchange of personally identifiable
care professions. We discuss these in turn. health care information by authorized individuals. Such pol-
icy could allow explicit recognition of the multiple uses of
The Medicare Modernization Act22 (MMA) supports pro-
medical record data within health care institutions and pro-
vider adoption of electronic prescribing technology and pro-
viders’ offices for billing, documentation, decision support,
vides for some relaxation of the Stark regulations in the Social
and quality analysis, as well as the patient’s rights with re-
Security Act (Section 1877). It is critical that these regulations
spect to secondary uses of the data beyond health care oper-
be supported in practice from two perspectives. First, physi-
cians in distinct organizational entities (different businesses) ations.
must be allowed to form purchasing cooperatives to allow To derive value from HIT in clinical settings, two things must
economies of scale to accrue in HIT purchasing decisions. happen simultaneously: functionally rich EHRs supporting
This would allow providers to experience considerable sav- comprehensive patient data management, decision support,
ings when participating in volume purchase agreements and health care workflow must be adopted in acute and
with vendors. Secondly, larger hospitals and integrated deliv- chronic care settings,3 and these systems must share data
ery networks must be allowed to improve the ability of physi- with one another. That is, clinical information systems in dis-
cians using their office technology to interact with that parate health care business entities must exchange clinical in-
hospital or IDN clinical information systems for review of pa- formation on common patients for treatment purposes.2
tient care data. In addition, health care data from the provider Given the rational but myopic business perspective of most
offices should be made available to the hospital systems. In health care providers, we believe that to achieve regional
situations in which community providers have affiliations health care information exchange, an appropriate regional
with multiple inpatient care facilities, this ability is particu- authority must be established to guide development and im-
larly important for patient safety and quality of care—they plementation of data sharing policies and procedures among
must have a complete view of their patients’ health care providers and patients, legal frameworks, enabling technol-
data from wherever care is provided. The MMA requirement ogies (e.g., patient matching algorithms), and management
that the hospital or IDN data be made available to any pro- of shared expenses and financial benefits in a coherent and
vider in the community may only be made possible through sustainable business model. Such regional health care informa-
a community health information exchange. tion exchanges are under development in several areas,10,21,28
18 MIDDLETON ET AL., Accelerating U.S. EHR Adoption

and several legislative efforts support this notion,29 but it stand the potential of HIT, its use, and its limitations. This
would be useful to have in place federal guidelines, and seed may begin by expanding the curriculum devoted to clinical
money, that could be applied locally and regionally to ensure informatics in the health professions schools. Finally, an edu-
their success. cational campaign should also be directed at the executive
Finally, with the advent of ‘‘wired’’ clinical care environments suites of our health care enterprises, both large and small,
and their emerging interconnectivity, and an increasingly where the strategic and investment decisions are made about
mobile patient, we suggest that soon it will be advantageous the business of health care. Without leadership and commit-
for providers and their patients to have licensure in the health ment, whether it is the small office environment or the largest
professions be provided at the federal level. Providers should IDN, adoption of HIT will not proceed.
be able to act on behalf of their patients even remotely; for ex- Even if every physician, nurse, and hospital were committed
ample, when a patient is in another state, experiences a med- to adopting HIT, however, to facilitate rapid adoption, it is
ical problem, and communicates electronically with his or her critical that we engineer adoption strategies that scale. That
provider at home (who has access to both the local and re- is, every clinic and hospital environment must not be forced
mote health care data). Short of national licensure per se, re- to rediscover best practices for implementing HIT; there
laxation of state regulations to facilitate reciprocity of should be a National Resource Center for HIT that can be a
professional licensure between state agencies is a worthy first repository of best practices and expertise for HIT implemen-
step. Broadening the geographic scope of licensure in the tation to accelerate the process. The recently announced re-
heath professions will allow the development of regional quests for proposals from the AHRQ appear to be well
health care information exchanges that truly reflect ‘‘medical targeted to meet this need. In addition, we believe there
marketplaces’’ that may span across state boundaries. With should be a national repository that would make available
national licensure or improved reciprocity between states, clinical knowledge required for HIT adoption, be readable,
providers would be able to physically practice more readily and be encoded in a standardized manner, including items
in more than one state. More importantly, however, as health such as appropriate controlled terminology, standard code
care becomes more ‘‘wired,’’ providers will be able to seam- sets, care rules, alerts and reminders, order sets, documenta-
lessly collaborate across state lines, rendering opinions re- tion templates, and forms, so that each clinic and hospital
motely from the patient care site or remotely performing does not have to rediscover the best clinical knowledge for
critical interpretive duties such as reading radiology, nuclear, implementation within their chosen clinical systems. We be-
electrocardiographic, sonographic, and other image modalities lieve the absence of such a resource protracts the implemen-
and interpreting biomedical signals, which do not require tation of HIT, and in some settings, the absence of the
physical proximity to the patient. appropriate resources will make sophisticated decision sup-
port in clinical systems an unattainable goal. These resources
Educational, Marketing, and Supporting Activities
should be public–private collaboratives that serve the inter-
Achieving President Bush’s vision—that most Americans
ests of the HIT marketplace as well as the public and private
would have an electronic health record within ten years—will
purchasers of health care.
require an extraordinary effort. In addition to the recommen-
dations above, we feel there is need for an educational and Conclusion
marketing campaign not dissimilar to the public announce- There is growing support for the widespread adoption of
ments and efforts surrounding smoking cessation, drug EHR as a fundamental strategy to improve U.S. health care
abuse, obesity, accident prevention, and other campaigns in delivery, efficiency, quality, and safety. Despite considerable
the interest of the public’s health. At the 2004 ACMI evidence to support adoption of EHR, progress has been
Retreat, Kevin Johnson of Vanderbilt University suggested slow to date. We suggest that the current HIT marketplace
a campaign: ‘‘Got EHR?’’ has failed because of several factors, including misalignment
We suggest a three-pronged marketing and educational cam- of financial incentives, absence of a clear business case for
paign directed at consumers, health care professionals, and EHR adoption and for interoperability between EHR imple-
the executive suites of our provider organizations across the mentations, and incomplete specification and adoption of rel-
country. The public has heard the news from the IOM reports evant standards. To accelerate EHR adoption we believe
that made the front pages of local newspapers, describing a variety of stimuli are needed to align incentives, provide
medical error,30 poor quality of care, and the role of HIT,31 new incentives for adoption of interoperable EHRs, coordi-
but they do not yet generally perceive the risks of receiving nate and promote relevant standards, and educate the health
care from providers and hospitals that do not have HIT care community and consumers. This report describes our
with clinical decision support in place. It is often mistakenly recommendations (summarized in Table 1) in all of these
believed that HIT is already in use.32 Just as a consumer buy- areas.
ing a car today would never think of selecting one without
seatbelts, airbags, or other safety features, American con- References j
sumers should ask whether their personal physician and their
hospital have CPOE and EHR systems in place and in use. 1. Berner E, Detmer DE, Simborg D. Will the wave finally break?
A brief view of the adoption of electronic medical records in
Secondly, health care professionals may be increasingly aware
the United States. J Am Med Inform Assoc. 2005;12:3–7.
of the potential benefits of HIT, but they have not yet adopted 2. Ash JS, Bates DW. Factors and forces impacting EHR system
these technologies to any significant degree. In conjunction adoption: report of a 2004 ACMI discussion. J Am Med Inform
with alignment of incentives and reimbursement reforms Assoc. 2005;12:8–12.
we have described, an education campaign needs to be di- 3. Johnston J, Pan E, Walker JD, Bates DW, Middleton B. The Value
rected toward health care professionals to help them under- of Computerized Provider Order Entry in Ambulatory Settings.
Journal of the American Medical Informatics Association Volume 12 Number 1 Jan / Feb 2005 19

Boston, MA: Center for Information Technology Leadership; 18. Steen EB, Detmer D (eds). The Computer-based Patient Record:
2003. An Essential Technology for Healthcare. Washington DC: Insti-
4. Pan E, Johnston D, Walker JD, Adler-Milstein J, Bates DW, tute of Medicine, National Academy Press; 1991.
Middleton B. The Value of Healthcare Information Exchange 19. de Brantes F. Bridges to excellence: a program to start closing the
and Interoperability. Boston, MA: Center for Information Tech- quality chasm in healthcare. J Healthc Qual. 2003;25(2):2–11.
nology Leadership; 2004. 20. Birkmeyer JD, Dimick JB. Potential benefits of the new Leapfrog
5. Wang SJ, Middleton B, Prosser LA, et al. A cost-benefit analysis standards: effect of process and outcomes measures. Surgery.
of electronic medical records in primary care. Am J Med. 2003; 2004;135(6):569–75.
114(5):397–403. 21. Valdes I, Kibbe DC, Tolleson G. Barriers to proliferation of elec-
6. Birkmeyer CM, Lee J, Bates DW, Birkmeyer JD. Will electronic tronic medical records. Inform Prim Care. 2004;12(1):3–9.
order entry reduce health care costs? Eff Clin Pract. 2002;5(2): 22. Commission on Systemic Interoperability. Medicare Moderniza-
67–74. tion Act of 2004 (S.2343.IS). Available at: http://www.cms.
7. Zdon L, Middleton B. Ambulatory electronic records implemen- hhs.gov/medicarereform/. Accessed August 2, 2004.
tation cost benefit: an enterprise case study. Proc Hlth Infor 23. Department of Veterans Affairs Office of Information. VISTA
Mgmt Sys Soc 1999;4:97–117. Monograph: Veterans Health Information Systems and Technol-
8. Brailer DJ, Augustinos N, Evans LM, et al. Moving Toward Elec- ogy Architecture (VISTA). Available at: http://www1.va.gov/
tronic Health Information Exchange: Interim Report on the vista_monograph/. Accessed August 2, 2004.
Santa Barbara County Data Exchange. Sacramento: California 24. McDonald CJ, Overhage JM, Tierney WM. The Regenstrief Med-
Healthcare Foundation; 2003. ical Record System: a quarter century experience. Int J Med Inf.
9. Consolidated Health Informatics. US Federal Government G2B 1999;54(3):225–53.
Porfolio. Available at: http://www.whitehouse.gov/omb/egov/ 25. Middleton B. Comments to the NCVHS on patient medical
gtob/health_informatics.htm. Accessed July 30, 2004. record information standards. Available at: http://ncvhs.
10. Overhage JM, Rischel W. Executive Summary. Technical Panel hhs.gov/991014tr.htm. Accessed August 2, 2004.
Recommendations: Achieving Electronic Connectivity in Health- 26. Tessier C, Waegemann CP. The continuity of care record. Organ-
care. New York, NY: Connecting for Health, Markle Foundation; izations unite to create a portable, electronic record for use at
2004. transfer points in patients’ care. Healthc Inform. 2003;20(10):54–6.
11. Radiological Society of North America, and the Healthcare Infor- 27. HL7. Electronic health record functional model and standard.
mation Management and Systems Society. Integrating the Available at http://www.hl7.org/ehr/. Accessed August 2,
Healthcare Enterprise. Available at: http://www.himss.org/asp/ 2004.
issuesbytopic.asp?TopicID=11. Accessed August 2, 2004. 28. Patient Safety Institute. Economic value of a community clinical in-
12. Glaser JP, DeBor G, Stuntz L. The New England Healthcare EDI formation sharing network. Available at <http://www.ptsafety.
Network. J Healthc Inf Manag. 2003;17(4):42–50. org/resources/PSI_cost_savings.pdf>. Accessed August 2, 2004.
13. Bates DW, Gawande AA. Improving safety with information 29. Kennedy P. Quality, efficiency, safety, and technology for health-
technology. N Engl J Med. 2003;348:2526–34. care transformation act. Available at <http://thomas.loc.
14. Bates DW, Pappius E, Kuperman GJ, et al. Using information gov/cgi-bin/query/D?r108:1:./temp/;r108gweoxT>. Accessed
systems to measure and improve quality. Int J Med Inf. 1999; August 2, 2004.
53(2-3):115–24. 30. Kohn LT. To Err is Human: Building a Safer Health System.
15. Henry SB, Lenert L, Middleton B, Partridge R. Linking process Washington, DC: National Academy Press; 2000.
and outcome with an integrated clinical information manage- 31. Committee on Quality of Health Care in America. Crossing the
ment system knowledge systems. Proc HIMSS. 1993:57–81. Quality Chasm: A New Health System for the 21st Century.
16. Poon EG, Wald J, Bates DW, et al. Supporting patient care Washington, D.C.: National Academy Press; 2001.
beyond the clinical encounter: three informatics innovations 32. Work Group on Policies for Electronic Information Sharing
from partners healthcare. Proc AMIA Symp. 2003:1072. Between Doctors and Patients. Connecting Americans to Their
17. Welsh F. Informatics: a physician’s view. J Health Care Finance. Healthcare. New York, NY: Connecting for Health, Markle Foun-
1997;23(3):37–43. dation; 2004.